The Yin and Yang of Pioneer Transcription Factors: Dual Roles in Repression and Activation

Abstract

Pioneer transcription factors, by virtue of their ability to target nucleosomal DNA in silent chromatin, play crucial roles in eliciting cell fate decisions during development and cellular reprogramming. In addition to their well-established role in chromatin opening to activate gene expression programs, recent studies have demonstrated that pioneer factors have the complementary function of being able to silence the starting cell identity through targeted chromatin repression. Given recent discoveries regarding the repressive aspect of pioneer function, we discuss the basis by which pioneer factors can suppress alternative lineage programs in the context of cell fate control.

Graphical abstract

The dual role of a pioneer transcription factor has emerged. Over the last several years, multiple studies have revealed that, upon its expression, a pioneer factor represses genes associated with a starting cell type. Then, it executes its canonical role, namely activating the expression of genes associated with a new cell type, which are originally embedded in closed chromatin.

INTRODUCTION

Pioneer transcription factors (TFs) play a crucial role in cell fate decisions during embryonic development and cellular reprogramming via their unique ability to enact persistent epigenomic changes^[1,2]. The name “pioneer factor” derives from an ability to facilitate chromatin opening; whereas most TFs can bind to regions of chromatin that are already accessible, pioneer factors have an increased ability to bind to DNA regions that are embedded in nucleosomes (“closed chromatin”) compared to other transcription factors^[3–9] and thereby increase chromatin accessibility to other TFs. Pioneer activity is especially evident in enhancer regions that are responsible for tissue-specific gene regulation. Consequently, gene induction by pioneer factors is critical for the durable changes in cell state or cell identity that occur during development, tissue metaplasia, and reprogramming^[10].

However, there is mounting evidence to suggest that pioneer factors also possess gene silencing activity. Nearly two decades ago, our group reported that FOXA1/2 pioneer factors interact with Groucho-related (Gro/TLE/Grg) corepressors^[11]. This leads to a FOXA-dependent repression of target genes due to a compaction of nucleosomal structure (i.e. chromatin closing), both in vivo and reconstituted with nucleosome array templates in vitro^[11]. Ten years later, Chronis et al. found that during the generation of induced pluripotent stem cells (iPSCs), the pioneer factors OCT4, SOX2 and KLF4 (OSK) exhibit dual activities, binding and activating the enhancers of pluripotency-associated genes while silencing the somatic enhancers associated with the starting cell identity^[12]. These seminal studies established a role for pioneer factor-mediated silencing in cell culture systems. Recently, this paradigm has been extended to the in vivo setting, where the importance of pioneer factor silencing activity has been demonstrated in several physiological processes, ranging from stem cell differentiation to cellular reprogramming^[13–16]. In this commentary, we take a closer look at the emerging literature on repression by pioneer factors and offer a new framework for understanding such in epigenetics and cellular identity. It will become clear from the discussion below that the various mechanisms for gene repression by pioneer factors are likely distinct from the nucleosome targeting that distinguishes pioneer factors with regard to gene activation. By highlighting these newly appreciated roles in gene repression, we hope to place pioneer factor function in a broader gene regulatory context.

PIONEER FACTOR-MEDIATED ENHANCER REORGANIZATION.

In 2017, Chronis et al. proposed the enhancer reorganization model in iPSC reprogramming, in which OSK factors bind and inactivate somatic enhancers early in reprogramming before exerting their more conventional pioneering activity^[12] (Figure 1A). The authors proposed three scenarios for decommissioning of the somatic enhancers. First, OSK binding at somatic enhancers led to an increase in histone deacetylase recruitment while histone acetylase p300 recruitment remained unchanged, thereby reducing H3K27ac levels in these regions. Second, OSK binding at somatic enhancers led to the removal of somatic TFs, including AP-1 family factors and CEBPα/β through an unknown mechanism. This, in turn, resulted in the redistribution of the somatic TFs to other sites that were already engaged by OSK, including pluripotency enhancers. Third, OSK expression downregulated transcription of the AP-1 factor genes, thereby reinforcing the loss of somatic enhancer activity. All these events occurred prior to activation of pluripotency enhancers targeted by OSK. Thus, according to this model, pioneer factor-mediated enhancer reorganization operates in a stepwise manner, with enhancer decommissioning preceding de novo activation.

Figure 1. Multiple scenarios proposed as the basis for pioneer factor-mediated silencing.

(A) Enhancer reorganization via redistribution of TFs present in the starting cells, as proposed by Chronis et al.^[20] and Thompson et al^[15]. Although not depicted in this diagram, Chronis et al. also confirmed recruitment of HDACs for removal of the active histone mark H3K27ac at the enhancers in the starting cells, as well as downregulation of somatic TFs, such as AP-1^[20].

(B) Pioneer factor-mediated redistribution of cooperative factors, which are required for the establishment and maintenance of active enhancers, including chromatin remodelers ARID1a/b and H3K4me1 writers MLL3/4, as reported by Yang et al^[14]. This report indicated that the pioneer factor SOX9 operates indirectly in this process, without targeting the active enhancers in the starting cells.

(C) Recruitment of corepressor complexes PRC1/2 can contribute to silencing enhancers^[15]. In this scenario, deposition of the repressive marks H2AK119ub and H3K27me3 and the active mark H3K4me1 (not depicted in the diagram) can occur simultaneously. This makes these enhancers transcriptionally poised, thereby preventing expression of precocious and alternative lineage genes.

(D) Our recent work revealed that the pioneer factor SOX4 can outcompete and evict hepatocyte master TFs such as HNF4A to initiate biliary reprogramming of hepatocytes in vivo^[16].

Five years later, in 2022, Thompson et al. reported that the competition for enhancers between the GATA6 pioneer factor and the pluripotency TF NANOG underlies the specification of lineages associated with the embryonic inner cell mass (ICM)^[13]. In normal embryos, ICM cells are bipotent in that they can either differentiate into pluripotent epiblast cells or primitive endoderm (PrE) cells, which form an epithelial layer that surrounds the epiblast^[13]. The authors found that mis-expression of GATA6 in pluripotent embryonic stem cells (ESCs), which most closely resemble the cells of the epiblast, induced a rapid and irreversible switch to the PrE state. During this cell fate switch, GATA6 bound to PrE enhancers to drive the new cell fate, as expected. But surprisingly, GATA6 was also found to bind epiblast enhancers. This binding was transient, but it was sufficient to evict pluripotency TFs, including NANOG and SOX2, thereby contributing to silencing of the epiblast phenotype. In turn, these evicted TFs became redistributed to regions of chromatin newly opened by GATA6 in a manner analogous to that described by Chronis et al. (Figure 1A).

PIONEER FACTOR-MEDIATED CHROMATIN CLOSING IN THE SKIN

The aforementioned cell culture-based studies established that, along with their well-recognized ability to open chromatin, pioneer factors also mediate cell state transitions via enhancer repression, including transcription factor eviction and chromatin closing. But whether these latter activities contribute to normal development or tissue homeostasis in vivo has only recently been addressed. In 2023, Yang et al. provided the first evidence for the silencing activity of a pioneer factor in animals by studying cell fate switching between mouse epidermal stem cells (EpdSCs) and hair follicle stem cells (HFSCs), a process driven by the master regulator pioneer factor SOX9^[14].

The authors began by confirming that ectopic expression of SOX9 in vivo converted EpdSCs to HFSC-like cells. Consistent with the sequential reprogramming mechanism engaged by OSK^[12] and GATA6^[13], silencing of the starting EpdSC phenotype preceded activation of the HFSC phenotype. In this case, however, the authors did not find evidence that SOX9 bound to the closing chromatin regions (Figure 1B), arguing against the possibility that SOX9 directly competed for binding with native EpdSC-associated TFs. Instead, they found that SOX9 competed for cooperative proteins responsible for chromatin opening, including chromatin remodelers (e.g., ARID1a, ARID1b), their AP-1 family co-factors (e.g., FOSL2, JUNB) and histone H3K4me1 modifiers (e.g., MLL3, MLL4) (Figure 1B). The competition was achieved remotely from the opening sites targeted by SOX9. Namely, redistribution to the newly opened regions decreased the relative availability of these cooperative proteins at the native enhancers, which were then decommissioned. Surprisingly, SOX9 maintained this competitive ability even without binding to DNA, as demonstrated by the maintenance of silencing activity by a mutant SOX9 with its HMG domain deleted. In contrast, deletion of the transactivating domain, which was required for recruitment of cofactors including chromatin remodelers, abrogated the silencing activity of SOX9 regardless of its DNA binding ability. Thus, the SOX9 pioneer factor contributes to chromatin closing during cellular reprogramming in the skin in vivo by mechanisms that may be distinct from those observed in vitro, at least in this cell context.

A PIONEER FACTOR ESTABLISHING BIVALENT ENHANCERS TO PREVENT ALTERNATIVE DIFFERENTIATION.

Earlier this year, Matsui et al. reported that a pioneer factor recruits corepressor complexes to prevent precocious and alternative-lineage gene expression using an in vitro endodermal development model^[15]. This demonstration that pioneer factors safeguard cells from deviating from their differentiation track offers new insight into how a pioneer factor can delay a developmental decision. The authors first discovered that knocking down all FOXA TFs (FOXA1, FOXA2, FOXA3) in human iPSCs surprisingly resulted in the induction of many lineage-unrelated genes when the cells were exposed to signals inducing endodermal differentiation. While the genes included those expressed at the late stage of endodermal differentiation, such as hepatic and pancreatic genes, they also included those expressed by mesodermal and ectodermal cells. This indicates that FOXA pioneer factors not only help induce endoderm but are also required to prevent precocious terminal differentiation or the erroneous acquisition of other lineage trajectories.

Mechanistically, FOXA pioneer transcription factors played a crucial role in establishing bivalent enhancers, which are important for regulating complex gene expression programs during development. The authors found that FOXA directly interacted with the transcription factor PR domain zinc finger 1 (PRDM1). PRDM1 serves as a hub for corepressor complexes, including the nucleosome remodeling and deacetylation (NuRD) complex (which catalyzes histone deacetylation) and Polycomb repressive complexes (PRCs) (which catalyzes the formation of H3K27me3 or H2AK119ub1) (Figure 1C). Interestingly, the FOXA-PRDM1-targeted regions were frequently marked by H3K4me1, an active enhancer mark, indicating that FOXA-PRDM1 cooperatively establishes bivalent enhancers. Notably, this silencing function was not restricted to FOXA pioneer factors, as the group also found that OCT4 directly interacts with PRDM14, thereby contributing to the establishment of bivalent enhancers in iPSCs.

PIONEER FACTOR COMPETITION WITH NATIVE TRANSCRIPTION FACTORS FOR BINDING MOTIFS.

Complementing these studies, we recently reported that the pioneer factor SOX4 initiates hepatocyte-to-biliary reprogramming in the mouse liver upon injury^[16]. Hepatocyte-to-biliary reprogramming, or biliary metaplasia, is a homeostatic response to liver injury whereby hepatocytes transdifferentiate into biliary epithelial cells (BECs)^[17–19]. We found that before exerting its canonical pioneer function to activate enhancers associated with biliary epithelial cells (BECs), SOX4 first silenced the original hepatocyte phenotype. Importantly, SOX4 mediated this silencing activity by competing with master regulator TFs responsible for hepatocyte identity, the starting cell state. Upon Sox4 expression, such TFs – including HNF4A, RXRA, and PPARA – were evicted from their canonical binding sites in hepatocyte enhancers. To our surprise, many of these hepatocyte TFs shared a core binding motif sequence with SOX4. For example, the SOX binding motif (CTTTGT/ACAAAG) overlaps with the HNF4A binding motif (CAAAG/CTTTG). Temporally, SOX4 bound first to HNF4A-targeted regions, evicting HNF4A from the engaged hepatocyte enhancers (Figure 1D). Remarkably, in vitro binding assays demonstrated that SOX4 protein has a stronger affinity for a canonical HNF4A binding site than HNF4A itself. These findings are consistent with the enhancer reorganization model established by Chronis et al.^[12] and Thompson et al.^[13] and provide a molecular explanation for the observed eviction of HNF4A and other hepatocyte-specific TFs by SOX4.

Competition for a shared binding motif suggests a novel framework for pioneer factor-mediated cellular plasticity. Motif competition by TFs is important in various biological contexts, including cell fate decisions during development and oncogenesis. For example, Hu et al. found that KLF4 and ZFP281 partially share a binding motif, which helps explain the transcriptional heterogeneity of mESC states^[20]. Our study extends this framework of motif competition, demonstrating that pioneer TFs can override pre-engaged active enhancers and directly initiate cell fate switching. Moreover, given that SOX4 has an HMG-box binding motif while HNF4A has a DR1 binding motif, our findings suggest that motif competition need not involve paralogous TFs, but can stretch across divergent TF families.

FURTHER THOUGHTS

What mechanisms account for pioneer factor repressive activity? As described above, at least 3 modes of action exist: (1) motif competition with native TFs in starting cells^[13,16], (2) recruitment of HDAC(s)^[12] or corepressor complexes including Gro/TLE/Grg^[11] and PRC1/2^[15], and (3) redistribution of (or competition for) transcriptional coactivators including histone modifiers^[12,14] and chromatin remodelers^[14]. Our recently published work^[16] supports the recruitment of Gro/TLE/Grg^[11] as an important feature of the silencing mechanism. Specifically, we observed upregulation of Foxa and Tle genes in Sox4-expressing hepatocytes during biliary reprogramming. Furthermore, the expression of these genes was also elevated in reprogrammed cells and native BECs compared to normal hepatocytes. These data suggest that Foxa and/or Tle contribute to the silencing of hepatocyte genes following SOX4 binding to hepatocyte enhancers. We therefore propose a scenario whereby eviction of HNF4A and other hepatocyte TFs by SOX4 is followed by the recruitment of FOXA/TLE, which reinforces the silencing of hepatocyte genes.

An important question regarding the outcome of pioneer factor binding remains unanswered: How do pioneer factors recruit coactivators in some contexts, while recruiting corepressors in others? Considering that the repressive activity of a pioneer factor is partly ascribable to either DNA-binding corepressors, such as PRDM^[15], or non-DNA binding corepressors, such as RFX factors^[21], the context in which a pioneer factor binds undoubtedly influences the characteristics of corepressor recruitment. Likewise, the frequency with which pioneer factors exert repressive activity remains unknown, given the limited number of relevant studies to date. While these issues are known to exist for many transcription factors capable of eliciting positive and negative regulatory outcomes, it is worth emphasizing that they also pertain to pioneer factors, given the longstanding assumption that these DNA binding proteins act in a strictly positive fashion.

CONCLUSION

Recent studies have greatly expanded our appreciation for the complex and multifaceted roles of pioneer factors in cellular reprogramming and differentiation. The recognition that pioneer factors play a repressive role in gene regulation represents a paradigm shift in our understanding of transcriptional regulation. Through various mechanisms, such as competition for binding motifs, recruitment of co-repressor complexes, and safeguarding against aberrant differentiation, these transcription factors orchestrate the intricate process of silencing a former cell identity while activating a new one. Further exploration into the context-dependent nature of pioneer factor activities and their cooperation with elements of the cellular machinery will provide deeper insights into the dynamic regulation of cellular plasticity and epigenetic control of cellular identity.

Data availability statement:

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.